The market for the use of recombinant protein therapeutics has increased steadily for the last quarter century. In 2005, six of the top 20 drugs were proteins, and overall, biopharmaceutical drugs accounted for revenues of approximately $40 billion, of which approximately $17 billion was based on the sales of monoclonal antibodies.
Monoclonal antibodies represent a distinct class of biotherapeutics with a great deal of promise. The antibody scaffold is well tolerated in the clinic, and glycosylated IgG molecules have favorable pharmacokinetic and pharmacodynamic properties. Comparison of the sequences of the approved antibody drugs, as well as those in development, demonstrates that some of the individual drug molecules are strikingly similar to each other, differing only by a few variations of amino acid residues located in the variable region of the immunoglobulin.
Typical monoclonal antibodies, like naturally occurring antibodies, have the appearance of a “Y”-shaped structure and the antigen binding portion being located at the end of both short arms of the Y. The typical antibody molecule consists of four polypeptides—two identical copies of a heavy (H) chain and two copies of a light (L) chain, forming a general formula H2 L2. It is known that each of the heavy chains contains one N-terminal variable (VH) plus three C-terminal constant (CH1, CH2 and CH3) regions and light chains contain one N-terminal variable (VL) and one C-terminal constant (CL) region each. The different variable and constant regions of either heavy or light chains are of roughly equal length (about 110 amino residues per region). Each light chain is linked to a heavy chain by disulphide bonds and the two heavy chains are linked to each other by disulphide bonds. Each heavy chain has at one end a variable domain followed by a number of constant domains, and each light chain has a variable domain at one end and a constant domain at the other end. The light chain variable domain is aligned with the variable domain of the heavy chain. The light chain constant domain is aligned with the first constant domain of the heavy chain. The remaining constant domains of the heavy chains are aligned with each other. The constant domains in the light and heavy chains are not involved directly in binding the antibody to the antigen.
Antibodies are typically divided into different classes on the basis of the structure of the constant region. In humans for example, five major structural classes can be identified immunoglobulin G or IgG, IgM, IgA, IgD and IgE. Each class is distinguished on the basis of its physical and biological characteristics which relate to the function of the immunoglobulin in the immune system. IgGs can be further divided into four subclasses: IgG1, IgG2, IgG3 and IgG4, based on differences in the heavy chain amino acid composition and in disulphide bridging, giving rise to differences in biological behavior. A description of the classes and subclasses is set out in “Essential Immunology” by Ivan Roitt, Blackwell Scientific Publications.
The variable domains of each pair of light and heavy chains form the antigen binding site. They have the same general structure with each domain comprising a framework of four regions, whose sequences are relatively conserved, connected by three complementarity determining regions (CDRs). The four framework regions (FWs or FRs) largely adopt a beta-sheet conformation and the CDRs form loops connecting, and in some cases comprising part of, the beta-sheet structure. The CDRs are held in close proximity by the framework regions and, with the CDRs from the other domain, contribute to the formation of the antigen binding site.
The vertebrate immune system has evolved unique genetic mechanisms that enable it to generate an almost unlimited number of different light and heavy chains in a remarkably economical way by joining separate gene segments together before they are transcribed. The antibody chains are encoded by genes at three separate loci on different chromosomes. One locus encodes the heavy chain isotypes and there are separate loci for the kappa (κ) and lambda (λ) light isotypic chains, although a B-lymphocyte only transcribes from one of these light chain loci. For each type of Ig chain—heavy chains, lambda (λ) light chains, and kappa (κ) light chain—there is a separate pool of gene segments from which a single peptide chain is eventually synthesized. Each pool is on a different chromosome and usually contains a large number of gene segments encoding the V region of an Ig chain and a smaller number of gene segments encoding the C region. More specifically, the variable region of an H-chain comprises three gene fragments, i.e., V, D and J gene fragments, while the variable region of an L-chain comprises two gene fragments, i.e., J and V gene fragments, regardless of whether the L-chain belongs to a lambda (λ) or kappa (κ) chain. During B cell development a complete coding sequence for each of the two Ig chains to be synthesized is assembled by site-specific genetic recombination, bringing together the entire coding sequences for a V region and the coding sequence for a C region.
The large number of inherited V, J and D gene segments available for encoding Ig chains makes a substantial contribution on its own to antibody diversity, but the combinatorial joining of these segments greatly increases this contribution. Further, imprecise joining of gene segments and somatic mutations introduced during the V-D-J segment joining at the pre-B cell stage greatly increases the diversity of the V regions
In addition to these structural characteristics, analyses of natural antibody sequences together with structural studies have been instrumental in revealing how antibodies work (Chothia et al., 1992, J. Mol. Biol., 227: 799-817; Kabat, 1982, Pharmacological Rev., 34: 23-38; Kabat, 1987, Sequences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, Md.)). These studies have shown that antigen recognition is primarily mediated by complementarity determining regions (CDRs) that are located at one end of the antibody variable domain and are connected by a β-sheet framework (Wu & Kabat, 1970, J. Exp. Med., 132: 211-250; Kabat & Wu, 1971, Annals New York Acad. Sci., 190: 382-393).
The sequence diversity of natural antibodies shows that the CDRs are hypervariable in comparison with the framework, and it is the CDR sequences that determine the antigen specificity of a particular antibody (Jones et al., 1986, Nature, 321: 522-5; Amit et al., 1986, Science, 233: 747-53). These studies have also revealed that the natural sequence diversity at most CDR positions is not completely random, as biases for particular amino acids occur in both a site-specific manner and in terms of overall CDR composition (Davies & Cohen, 1996, Proc. Natl. Acad. Sci. USA, 93: 7-12; Kabat et al., 1977, J. Biol. Chem., 252: 6609-16; Zemlin et al., 2003, J. Mol. Biol., 334: 733-49; Mian et al., 1991, J. Mol. Biol., 217: 133-51; Padlan, 1994, Mol. Immunol, 31: 169-217).
In contrast to traditional small molecule based approaches, therapeutic antibodies have significant advantages, including (i) their ability to be generated and validated quickly; (ii) therapeutic antibodies exhibit fewer side effects and have improved safety profiles, (iii) therapeutic antibodies have well understood pharmacokinetic characteristics, and they can be optimized to create long half-life products with reduced dosing frequency; iv) therapeutic antibodies are versatile and exhibit flexibility in drug function; v) therapeutic antibody scale-up and manufacturing processes are robust and well-understood; and vi) they have a proven track record of clinical and regulatory success.
Even given the success of monoclonal antibodies, the antibody-as-drug modality is continuing to evolve, and subject to inefficiency. Further, intrinsic biological bias within the native immune system often works against the more rapid development of improved therapeutics. These limitations include, i) the long development time for the isolation of biologically active antibodies with affinity constants of therapeutic caliber, ii) the inability to raise antibodies to certain classes of protein targets (intractable targets), and iii) the intrinsic affinity ceiling inherent in immune system based affinity selection.
Specifically there is a need for methods to more rapidly develop antibodies with improved pharmacokinetics, cross-reactivity, safety profiles and superior dosing regimens. Central to this need is the development of methods that enable the systematic analysis of potential epitopes with a protein, and enable the selective development of antibodies with the desired selectivity profiles.
An approach used by a number of companies includes the use of random or semi random mutagenesis (for example the use of error prone PCR), in conjunction with in vitro molecular evolution. This approach is based on the creation of random changes in protein structure and the generation of huge libraries of mutant polynucleotides that are subsequently screened for improved variants, usually through the expression of the encoded proteins within a living cell. From these libraries a few improved proteins may be selected for further optimization.
Such in vitro mutation approaches are generally limited by the inability to systematically search a significant fraction of sequence space, and by the relative difficulty of detecting very rare improvement mutants at heavy mutagenesis loads. This fundamental problem arises because the total number of possible mutants for a reasonably sized protein is massive. For example, a 100 amino acid protein has a potential diversity of 20100 different sequences of amino acids, while existing high throughput screening methodologies are typically limited to a maximum screening capacity of 107-108 samples per week. Additionally such approaches are relatively inefficient because of redundant codon usage, in which up to around 3100 of the nucleotide sequences possible for a 100 amino acid residue protein actually encode for the same amino acids and protein, (Gustafsson et al. (2004) Codon Bias and heterologous protein expression Trends. Biotech. 22 (7) 346-353).
A more sophisticated approach uses a mixture of random mutagenesis with recombination between protein domains in order to select for improved proteins (Stemmer Proc. Natl. Acad. Sci. (1994) 91 (22) 10747-51). This approach exploits natural design concepts inherent in protein structures across families of proteins, but again requires significant recombinant DNA manipulation and screening capacity of a large number of sequences to identify rare improvements. Both approaches require extensive follow-up mutagenesis and analysis to understand the significance of each mutation, and to identify the best combination of the many thousands or millions of mutants identified.